--- IO-AIO/Changes 2008/04/12 23:33:11 1.138 +++ IO-AIO/Changes 2010/01/10 20:37:33 1.221 @@ -3,15 +3,193 @@ TODO: better autoconf.pm that can return LIBS etc. TODO: aio_cptree/mvtree TODO: reduce condvar fairness: schedule hot-cache-threads first? -TODO: splice/tee/vmsplice/sync_file_range? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) +TODO: splice/tee/vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) TODO: aio_fcntl, at least for file-locking +TODO: fallocate, if it wouldn't be so useless. +TODO: IO::AIO::mmap? aio_mincore? + - (libeio) more fixes for the freebsd/apple sendfile - broken by + manpage standards, but apparently correct accoridng to actual + kernel sources. + - add IO::AIO::m(un)lockall support. + - clean up manpage. + +3.5 Thu Jan 7 21:25:04 CET 2010 + - (liebio) fix freebsd sendfile (Vanilla Hsu). + - (libeio) also fix darwin sendfile that suffered from a similar bug. + - add aio_statvfs as interface to statvfs/fstatvfs. + - work around buggy symlink() on cygwin in the testsuite. + - wtf. freebsd-8.0 actually passes the testsuite! I knew + they could implement fork and semaphores one day! cheers! + +3.4 Sat Jan 2 15:13:04 CET 2010 + - (libeio) max_poll_time was not properly converted to ticks. + - clarify cancel_subs description. + - IO::AIO::sendfile did not actually return the return value + from the sendfile call. + - implement aio_msync, aio_mtouch. + - (libeio) tentatively support darwin in sendfile. + +3.31 Thu Nov 12 02:14:29 CET 2009 + - fix result status documentation of aio_copy, aio_move. + - speed up object creation considerably by avoiding + hash lookups. + +3.3 Wed Aug 5 13:52:58 CEST 2009 + - use common::sense. + - use common schmorp.h header. + - allow integers (file descriptors) in addition to file handles. + - take advantage of linux' eventfd (but it seems slower on smp?). + - use poll not select on posix platforms. + +3.261 Wed Jul 1 10:11:51 CEST 2009 + - more 0S X workarounds (patch by Tokuhiro Matsuno). + +3.26 Tue Jun 30 09:33:26 CEST 2009 + - 0S X of course claims to be posix 2008 but lacks posix_fadvise. + try to work around this horribly broken OS in a somewhat hackish + way. might help other os's too. + +3.25 Sat Jun 27 05:18:26 CEST 2009 + - added IO::AIO::fadvise and IO::AIO::sendfile. + - (libeio) replaced quicksort+insertion sort by a tuned radix + sort + insertion sort, resulting in comparable runtime (usually + faster) to the old version, but without any ill side effects on + degenerated (for quicksort) data. + - (libeio) correctly sort dirs by inodes if we have perfect type + knowledge and DIRS_FIRST is used. + - (libeio) internally the wrong DT_* constants were used, but no + known system has differing EIO_DT_*/DT_* values, so not an issue. + - removed a (harmless) assert that was left over on the code + but should not have been. + - use more correct types (change IO length from IV to VAL64, + which makes a difference on systems without 64 bit perls). + +3.23 Sat Jun 13 16:57:58 CEST 2009 + - fix off-by-one bug in aio_readdir that was introduced in the + rushed 3.22 release. + +3.22 Sat Jun 13 15:32:40 CEST 2009 + - speed up readdirx sort algorithm slightly. + - bin/treescan was missing from distro tarball. + +3.21 Fri Jun 12 18:45:53 CEST 2009 + - new options --dirs and --files for treescan. + - install bin/treescan by default. + - (libeio) aio_readdir can now be cancelled while executing. + - fix a printf format string for 64 bit systems (could lead + to problems on big endian 64 bit systems). + - do not use qsort() but our own algorithm: glibc initialises + the whole locale and I/O subsystem inside qsort, causing + a stack overflow on 32 bit machines. The new sort uses much less + stack and is more than twice as fast in typical situations. + +3.2 Sun Jun 7 20:30:05 CEST 2009 + - (libeio) pwrite emulation was even more flawed than intended and did + not restore the file offset. + - add aio_readdirx, which can return inode and filetype and sort + the names in various ways. + - unfortunately, utime, chmod, chown on an open file that has just + been written can easily block, which caused aio_copy to block the + process. no more! + - no longer rely on dst path in aio_copy when futime is available. + +3.19 Tue Apr 21 22:05:21 CEST 2009 + - more perl 5.10 workarounds for aio_read and write. + - aio_write no longer modifies the sv (if possible). + - aio_read now works correctly with magic values. + +3.18 Sun Apr 19 21:17:32 CEST 2009 + - better diagnostics when some aio_* functions get passed an illegal fh. + - try to avoid crashes due to incompatible 5.10 API changes. grmbl. + +3.17 Thu Nov 20 08:45:36 CET 2008 + - (libeio) added aio_sync_file_range (untested). + - add aio_busy to @IO::AIO::AIO_REQ. + +3.16 Wed Oct 22 18:28:01 CEST 2008 + - use SvREFCNT_inc instead of SvREFCNT_inc_NN in a non-speed critical + part to improve portability to perl 5.8 (reported by szymon). + +3.15 Mon Oct 13 00:39:55 CEST 2008 + - automatic removal of feeders was broken. + - (libeio) use a more robust method to detect whether a feeder + has added something to the group or not. + +3.1 Thu Oct 2 13:34:40 CEST 2008 + - pre-deref the passed callback object, for increased speed + and decreased memory usage. + - call on_next_submit callback even for aio_group - impact unknown, + but seems more correct. + - $req->cb now returns the original callback. + - (libeio) pass EIO_GROUP/aio_group requests through the queue without + requiring a thread switch. + - (libeio) status code of aio_futime and aio_utime was always passed 0. + - do some other ยต-optimisations. + +3.07 Sat Aug 2 16:06:13 CEST 2008 + - do not include LIBS in autoconf tests. + +3.06 Tue Jul 15 12:41:32 CEST 2008 + - move preadwritelock definition before it's first reference, + to make it compile again on broken platforms (they still exist) + or badly cofngiured perls (redhat...), reported by Rob Bloodgood. + +3.05 Thu Jun 19 23:23:52 CEST 2008 + - work around perl overriding readdir etc. with thread-unsafe + (sic!) versions. + +3.04 Wed Jun 18 01:35:38 CEST 2008 + - (libeio) fix eio_mknod, which confused it's arguments. + - (libeio) do not use readdir_r, as Ulrich Drepper pointed out that this + is stupid. + - (libeio) fix eio__readahead prototype, patch by Jost Krieger. + - (libeio) fix a bug that could potentially cause IO::AIO + not to get initialised properly. + +3.03 Thu May 29 05:33:30 CEST 2008 + - (libeio) correctly call pthread_attr_destroy (). + - (libeio) work around broken bsd headers once more. + - reduce shared library size again by not including wrappers. + - max_outstanding could cause poll to enter a busy-waiting loop. + - document the new IO::AIO::poll_cb result value. + +3.02 Mon May 12 02:32:02 CEST 2008 + - fix a memory leak on aio_readlink. + - bring back working fchmod. + - nop and busy now set result to 0. + - set errno to ENOMEM when allocation fails. + +3.01 Sun May 11 03:07:03 CEST 2008 + - (libeio) make it compile on systems without readahead or readdir_r. + - (libeio) improve configure check for readahead availability. + - do not try to link against -lrt. + - use a separate configure script for IO::AIO (not the libeio one). + +3.0 Sun May 11 00:57:14 CEST 2008 + - added bin/treescan to the distribution. + - switched to using libeio. + - LOTS OF INCOMPATIBLE CHANGES: + - remove signal functionality, it is not worth the effort. + - max_outstanding no longer returns the previous number of requests. + - poll_cb no longer returns number of requests. + +2.62 Sat Apr 26 13:59:33 CEST 2008 + - port to solaris perls configured for old posix + (analysed by Jost Krieger). + - keep a reference to the perl filehandle object in aio_close, + so it doesn't get closed prematurely, leading to ugly races. + +2.61 Wed Apr 16 18:45:02 CEST 2008 - fix treescan output duplication and improve output ordering. also display files given on the commandline. + - use a different algorithm for aio_close that is faster + and probably has even lower chances of blocking. - do our own stack memory management for threads - linux allocates outrageous amounts of VM (not so bad unless you use mlockall...), which severely limits the number of threads - on 32-bit arches. + on 32-bit arches: stack size is the larger of PTHREAD_STACK_MIN + and 4096 * sizeof (long) (usually this is 16..64k). 2.6 Sun Mar 30 08:28:11 CEST 2008 - added aio_sync.